Why Do We Have Nightmares?

The dream starts normally enough. Then something is wrong. The setting warps, the situation turns dangerous, and you are running, or falling, or trying to scream but can’t. You wake up with your heart rate elevated, sometimes gasping.

Most people have nightmares. They are more common in children (who average one a week) and decline with age, though they persist throughout life. A subset of adults — roughly 4–8% — experience nightmares severe enough to cause sleep avoidance, daytime distress, and interference with functioning. This is called nightmare disorder.

But occasional nightmares, the kind almost everyone experiences, appear to serve a purpose.

The Threat Simulation Theory

The leading evolutionary account comes from cognitive neuroscientist Antti Revonsuo at the University of Turku in Finland. In 2000, he proposed the threat simulation theory: dreams, and nightmares specifically, are a biological mechanism for rehearsing responses to threatening situations.

The argument: the brain during REM sleep has access to threat-relevant emotional and memory systems — the amygdala is highly active, threat-processing circuitry is online, but the motor system is suppressed (the same REM atonia that prevents sleepwalking). In this state, the brain can simulate threatening scenarios without any risk of real consequences.

Running from a predator in a dream. Being trapped, attacked, socially humiliated, or lost — these common nightmare categories map directly onto the threat scenarios that mattered to human ancestors. The simulation theory argues the brain rehearses these scenarios specifically, not at random, because practicing threat responses while offline is adaptive: you become better at recognizing and responding to threats without having to experience them in reality.

Evidence for the theory: nightmares are disproportionately populated with threat scenarios. They rarely involve the experience of boredom, routine, or minor daily events. When researchers catalog nightmare content across cultures, the same threatening categories appear: being chased, falling, being unable to move or scream, confronting an intruder. These are survival-relevant themes, not random sampling of daily experience.

REM Sleep as Emotional Processing

Matthew Walker’s research on sleep and memory adds a related dimension.

REM sleep appears to perform a specific function: it reactivates emotional memories from the day and processes them in a neurochemical environment that is different from waking. During REM, norepinephrine — a stress-related neurochemical — is almost entirely absent from the brain. The emotional content of memories can be replayed in this norepinephrine-free state, which Walker hypothesizes allows the emotional charge to be processed and reduced without re-triggering the full physiological stress response.

On this account, REM sleep doesn’t just consolidate memories — it does emotional editing, separating the memory from the acute distress that originally accompanied it. The theory is sometimes called the “overnight therapy” hypothesis.

Nightmares may be a version of this process applied specifically to high-threat emotional content. The brain is working through something difficult, running it again in the relatively safe offline state, attempting to reduce its emotional charge.

This explains why nightmares often cluster around stressful periods — the emotional processing demands are higher, so the dream content reflects what’s being worked on.

When the Process Gets Stuck

The adaptive account works when the nightmare accomplishes something — when the emotional processing reduces distress and the threat content fades over time.

In PTSD, this process appears to fail. Trauma survivors often have nightmares that replay the traumatic event with high fidelity — not the kind of loose, associative dream typical of normal REM processing, but a near-literal rerun. These nightmares don’t fade. They can persist for years, waking the person with the same physiological terror response each time.

Walker’s hypothesis about why this happens: the norepinephrine suppression that normally allows REM’s emotional editing fails in PTSD. In people with PTSD, norepinephrine levels remain elevated during REM, preventing the emotional separation that normally occurs. The brain replays the trauma but can’t process it — the horror is re-experienced rather than reduced.

This has therapeutic implications: prazosin, a drug that blocks norepinephrine receptors, has been shown to reduce PTSD nightmares significantly in controlled trials — the closest thing to direct pharmacological evidence for Walker’s mechanism.

Why You Wake Up at the Worst Moment

One widely reported feature of nightmares: they often reach maximum intensity just before waking. The chase is about to end badly; the fall is nearly complete; the threat is at its most acute — and then you’re awake.

This may not be coincidence. The amygdala activation during intense nightmare content may itself trigger the arousal system, waking the sleeper at the moment of peak emotional intensity. In terms of the threat simulation theory, this is the equivalent of a fire drill ending before the simulated fire spreads — the practice has run long enough to be useful, and the safety system terminates it.

Whether this is a designed termination or simply the consequence of the amygdala alarm becoming loud enough to wake you is unclear.

The Content Patterns

Cross-cultural nightmare research by Michael Schredl and colleagues finds consistent content across cultures:

• Being chased or attacked: the most common category globally
• Falling: near-universal
• Being unable to move, scream, or act effectively: also universal (and overlapping with sleep paralysis experiences)
• The death of someone close: more common in adults than children
• Social failure, humiliation, or rejection: more prominent in cultures with high social anxiety norms

The consistency of content across disparate cultures is the strongest argument for evolutionary underpinning — these themes weren’t transmitted culturally, they appear wherever human brains run dreams.

The nightmare has a function. It is practice.

The problem is that practice can go wrong — can loop, can replay without resolution, can become the thing you’re practicing to avoid.

Most nights, the simulation runs, the threat is processed, and you wake up slightly more prepared for something you may never face.

Occasionally, the drill becomes the disaster.

That’s nightmare disorder, and it’s the same system working at full volume with nowhere to put what it’s holding.